Co-registration of mass spectrometry and optical whole slide histology images for computational pathology

Matrix-assisted laser desorption ionization mass spectrometry imaging (MALDI-MSI) enables interrogating the molecular composition of tissue with ever-increasing resolution and sensitivity. The output of a MALDI-MSI analysis is a sparse 3D tensor depicting the spatial distribution and relative ion intensity of molecular features differentiated by mass-to-charge ratio. Optical whole slide images (WSIs) can provide morphological characteristics of the same tissue section, due to the nondestructive nature of MALDI-MSI. The bimodal analysis of the morpho-molecular signal might enhance downstream diagnostic tasks and offer crucial support to pathologists [1, 2]. We propose an operator-free and data-efficient co-registration toolbox using the ANTsX ecosystem [3]. We develop and validate an intensity-based method, inspired by the work of W.M. Abdelmoula et al. [4], to estimate the affine transform parameters linking the coordinate systems of the two spatially-resolved modalities: the MALDI-MSI (.imzML) and the WSI (e.g., .tif, .svs); additionally, we also linked them to the SCiLS Lab closed-source software (Bruker Daltonics, Germany). We tested the method on more than 50 image pairs from FFPE tissue sections, comprising lung and thyroid tissue microarrays (TMAs), kidney and thyroid resections, and organoids. The molecular data is acquired with tims TOF fleX and rapifleX instruments at a spatial resolution of 20 and 50 μm. Compared to manual coregistration, in two thyroid TMAs of 86 patients, we achieved higher cell coverage (95 wins out of 132 cores) within histone-expressing regions and higher connective cells coverage (78 wins out of 118 cores) within collagen-expressing regions. References: [2] B. Balluff, et al. An overview of image registration for aligning mass spectrometry imaging with clinically relevant imaging modalities. JMSACL, 23 (2022) 26-38; [1] V. Coelho, et al., Improving the Annotation for Spatial Proteomics: A Computational Approach to Enhance Molecular Characterization of Thyroid Nodules, J. Proteome Res., (2026); [3] https://github.com/ANTsX/ANTs; [4] W.M. Abdelmoula, et al. Automatic Generic Registration of Mass Spectrometry Imaging Data to Histology Using Nonlinear Stochastic Embedding. Anal. Chem., 86 (2014), 9204-9211. Acknowledgements: The proteomics and metabolomics unit at UniMiB, guided by Prof. Fulvio Magni. The cancer molecular pathology unit at Fondazione IRCCS San Gerardo dei Tintori, guided by Prof. Fabio Pagni. Funding: Fondazione Cariplo (2023-1804).